Review with evidence
Assess AI answers, generated data, code, or agent actions against clear criteria. Identify factual mistakes, missing context, unsupported claims, and meaningful failures.
Careers / Expert AI evaluators
Good judgment.
Better AI.
Join Rubrex as an expert evaluator. Bring the knowledge to spot what an AI system gets wrong, explain why it matters, and help define what good looks like.
01The work
Explain the failure.
Help set the standard.
Rubrex is an AI evaluation and engineering firm. Expert evaluation connects model behavior to the requirements of real tasks.
Make quality measurable
Help develop representative cases and scoring rubrics. Explain the difference between a pass, a partial result, and a failure using examples others can inspect.
Calibrate and investigate
Compare judgments, discuss ambiguous cases, and refine the criteria. Document disagreement and uncertainty instead of hiding them inside a score.
02Bring a specialty
Depth in your field.
Clarity in your reasoning.
Tell us where your experience is strongest. These are areas of interest, not a list of confirmed vacancies.
Software and technical systems
Code quality, debugging, data workflows, tool use, and the behavior of software agents.
Research and subject expertise
A specialist field where you can assess correctness, source quality, assumptions, and the limits of an answer.
Language and communication
Writing quality, instruction following, multilingual behavior, clarity, and suitability for the intended reader.
Security and data practices
Relevant experience with application permissions, AI security, data handling, or privacy evidence. Assignments depend on demonstrated expertise and scope.
03What makes a strong evaluator
A score needs
a reason behind it.
Show relevant experience, careful reading, and the ability to explain a decision using evidence. Strong evaluators follow a rubric consistently, identify when it is ambiguous, and revise a judgment when the evidence changes.
We value clear written reasoning, dependable communication, and respect for confidential information. Share only work you are authorized to show.
See how we think about rubrics04Before you introduce yourself
A few useful answers.
Do I need previous AI evaluation experience?
Relevant expertise and the ability to explain your judgment are a useful starting point. Tell us about any reviewing, research, testing, annotation, or evaluation experience. Requirements depend on the assignment.
Is this a specific job opening?
This page collects expressions of interest in expert evaluation work. It does not advertise a particular vacancy or guarantee employment or an assignment. We contact applicants when their experience may fit an evaluation need.
What are the location, schedule, and compensation terms?
These depend on the engagement. Include your availability and, if useful, your time zone. Location requirements, scope, compensation, and engagement terms are discussed and agreed before any assignment.
What should I send?
A concise description of your specialty, relevant experience, and availability. You may include a public profile or a work sample you have permission to share. A company email or file upload is not required.
How is my introduction handled?
Your name, email, professional introduction, and any links you include are stored in our Notion database for the Rubrex team to review. We use your contact details to discuss potential evaluation opportunities. The contact-data page explains the form and how to request correction or removal.
05Introduce yourself
Bring your expertise.
Show your reasoning.
Tell us your area of expertise, relevant experience, and availability. Include an optional public work sample or profile link. We’ll contact you if there is a potential fit.
evals@rubrex.aiWHAT HAPPENS NEXT
- We review your introduction against evaluation needs.
- If there is a fit, we discuss your experience and approach.
- Scope, compensation, and engagement terms are agreed before any assignment.